7.6 Operation for preventing failures caused by heavy load
To prevent slowdown of servers and hosts due to heavy load, check whether the settings in definition files and the system configuration are appropriate based on the following information:
-
Information about the OS
-
Information about the programs that have an interface with HA Monitor
-
Slowdown information collected in the monitoring history file
The monitoring history is a record of slowdowns that have occurred in servers and hosts.
HA Monitor collects monitoring history for servers running on the local host in the server mode and the local and remote hosts. This monitoring history is output to monitoring history files. The monitoring history files contain information about slowdowns that exceed the server failure monitoring time and host failure monitoring time. They also contain information about slowdowns that do not reach the server failure monitoring time or host failure monitoring time. In other words, slowdowns that HA Monitor does not detect as failures are also recorded in a monitoring history file. Analyzing the history of slowdowns that are not detected as failures enables you to do the following:
-
Reevaluate the values set in the definition files for the server failure monitoring time and host failure monitoring time.
-
Reevaluate the system configuration based on the actual system load status.
If you create a shell beforehand, you can use it to investigate the cause of server slowdowns when slowdowns occur.
This section explains how to collect and analyze monitoring history and how to investigate the causes of server slowdowns.
- Important
-
Server monitoring history cannot be collected in the following cases:
-
The server is a standby server.
Collection of server monitoring history and detection of server slowdowns are not supported in the following cases:
-
The server is running in the monitor mode.
-